Pure & Appi. Chem., Vol.55, No.5, pp.833—844, 1983 0033—4545/83/050833—12$03.OOPrinted in Great Britain. Pergamon Press Ltd

©1983 IUPAC

THE NATURE OF DEFECTS AND THEIR ROLE IN LARGE DEFORMATION ANDFRACTURE OF ENGINEERING THERMOPLASTICS

H.H. Kausch

Polymer Laboratory, Swiss Federal Institute of TechnologyLausanne, 32, chemin de Bellerive, CH—1007 Lausanne

Abstract — In this paper molecular defects (chain scission, weakbonds), network irregularities (mostly concerning entangle—ments), flaws and inclusions and production related defects arediscussed with respect to crazing, large deformations and rup—ture of Dolymers.

I NTRODUCTI ON

Plastic materials are made up of long chain molecules which are intercon—

nected by chemical crosslinking points (thermosetting resins, vulcanizedrubbers), crystalline regions (semi—crystalline thermoplastics), and physi—cal entanglements (practically all polymers). The networks formed in thisway are often modified (stabilized, plasticized, filled and/or reinforc—

ed). An average commercial plastic material, therefore, is a heterogeneoussystem. Even if it is perfect, it contains quite differer't structural ele-ments. Depending on their mode of interaction (slippage or elastic load-

ing), a variety of responses occurs (e.g. viscoelastic deformation, craz-ing, yielding, brittle or ductile failure). In a previous presentation (1)the molecular mechanisms leading to different failure patterns have beendiscussed. In this paper prime consideration will be given to material de-fects on a molecular, microscopic and macroscopic level. In order to eluci-date their role, the "normal behavior" of stressed polymers will briefly beintroduced in the following section.

STRUCTURAL ELEMENTS AND THEIR RESPONSE TO LOAD

In a defect-free amorphous polymer such as PS, PMMA, or PC one would haveto consider as structural elements the chain ends, the statistical chainsegments (0.4 to 1 nm in length), the chain sections between entanglementpoints (2 to 10 nm), and the statistically coiled molecules. Depending onthe production technique, a distinct globular superstructure may also bepresent (as for instance in PVC). In the case of crystalline polymers evi-dently the presence of lamellar crystalline regions and of spherulites hasto be taken into account.

The normal response of a thermoplastic material to increasing loads will bean anelastic widening of the "lattice", bond rotation in the main chainleading to a change of conformation and creep, segment rotation and crazeinitiation (where appropriate), and a brittle (fracture) or ductile insta-

bility (segment slip and yielding followed by large scale chain orientationand either strain hardening or flow). In the presence of a superstructurethese phenomena will preferentially develop at the intergranular or inter-spherul itic boundaries.

An apparent paradox should be noted. Whereas it has become quite clear thatchain scission plays no significant role in the deformation and fracture of

isotropic polymers (2), chain length is a quite important parameter. Thisis well demonstrated by a plot of fracture surface energy GIc vs molecu-

lar weight (Fig. 1).

Short chains with a molecular weight smaller than the entanglement molecu-lar weight Mc are easily - and practically individually - seperated, thusthe fracture surface energy Gic is of the order of the surface work para-meter y. With growing chain length entanglements are being formed whichestablish a physical network. At I < Tg entanglements are stable. If,therefore, in the course of a fracture process a chain has to be seperatedfrom such a network, this can only be done after considerable deformationof several chains, i.e. after a certain plastic deformation. Thus, fractureenergy and generally strength and thoughness as well increase with thevolume effected by the plastic deformation.

833

Fig. 1. Fracture surface energy

(Gic or G10) vs molecular weightfor PNMA (own measurements) andPS (literature values).

From these remarks it can be deduced that a plastic material will mostlikely be damaged if the molecular chains are degraded or weakened, if thenetwork is disentangled, plasticized or embrittled, and if local stressconcentration occurs through inclusions, flaws or internal tensions.

In the following a number of examples for each of these defects will be

given.

STRENGTH EFFECTS OF MOLECULAR AND STRUCTURAL DEFECTS

Chain degradationThe principal consequences of production or service related chain degrada-tion are the reduction of molecular weight, plasticization through monomerformation, and/or embrittlement through crosslinking reactions or chainweakening. There are several potential mechanisms for the destruction oflong chain molecules as for instance mechanical rupture (2), thermal decom-

position and UV degradation. A given polymer is generally not equally sus-ceptible to all types of destruction since the mechanisms are different.The polysulfone (PSU) is an example in cause. Thus, it shows excellent me-chancal strength and thermal resistance at temperatures up to and beyond160 C. If exposed to UV irradiation, however, it degrades fairly rapidly.In the latter process main chain scission occurs at the phenyl-C, phenyi-Oand phenyl-S bonds (3), side group reactions involve the CH3-groups. Theformed free radicals react with oxygen leading to the appearance of new,oxygen containing groups such as —OH, —CHO or -COOH (Fig. 2). The molecularweight decreases notably during such an irradiation: from Mw 72 000 attirr = 0 to M = 19 000 at tirr = 75 h. Concurrently the tensilestrength and the extensibility of the irradiated film samples decrease

(Fig. 3).

Fig. 2. Part of the IR spectrum of UV irradiated poly-sulfone films of 17 pxn thickness; tirr indicates expo-sure time to accelerated aging in Xenotest 150; notabledecreases of transmission in the region of 0-H stretchand C=O stretch vibrations.

Fig. 3. Decrease of uniaxial strength b of 17 xm thickpolysulfone film strips exposed to accelerated aging inXenotest 150 (o indicate point of rupture).

834 H. H. KAUSCH

102

:1010

/0/,

80

60

40

30

b I MPct

tirr Oh

Psu

3500 3000 2500

10

Fig. 2

tirr 52h0

2000 1800 1600 0 1 2 3 4 5 6 7

7Icm-1 Cb//.Fig. 3

Nature of defects in engineering thermoplastics 835

An almost opposite behavior is shown by polymethylmethacrylate (PMMA). Ir-radiation by UV does not measurably influence the molecular weight and themechanical properties. However, it is well known that PMMA is susceptibleto depoymerization at high temperatures (T > 200 °C). The formed monomermay act as plasticizer and reduce yield and rupture stresses at tempera—tures just below Tg•

The influence of thermal oxidation on the mechanical behavior of polymershas been known for a long time. Understandably, processing is very oftenthe most critical moment because of the necessarily higher extrusion ormolding temperatures. If, for instance, the extrusion conditions of poly-ethylene pipes are not well controlled, layers of oxidized material areformed at the inside pipe surfaces (4—6). These layers, which are verybrittle, have been studied in detail by Gedde, Jansson and Terselius(4-5). Two mechanisms contribute to an embrittlement of the material: theweakening of the backbone chain by formation of carbonyl groups and thecrosslinking of neighboring chains by C-O-C bridges. The authors identifiedboth types of bonds (-C=O; -C-U-C-) by IR reflectance spectroscopy at theinside walls of some HOPE pipes extruded at 200 °C. The thoroughly oxidizedlayers exterded to a depth of some 20 to 160 tm. These layers were extreme-ly brittle and cracked upon drawing at strains of a few per cent. The timesto fracture in static loadinq of internally pressurized pipes (average hoop

stress & = 4.2 MPa, T = 80 °C) and of uniaxial creep samples, cut fromthe wall, were considerably shorter in the case of oxidized material. Thefinal cracks seemed to originate at a distance of 0.2 to 0.3 mm from the

oxidized layer (4).

On the basis of these investigations it may be suggested that one can dis—tinguish in the oxidized HDPE pipes four different layers (distances s arewith respect to the inner wall):

5 = 0 - 100 pin thoroughly degraded, brittle material, t() 0s = 150 — 250 pm somewhat degraded, readily creeping material, ot(S)

smalls = 300 - 500 pm slightly damaged zone, ot(s)s > 500 tm normal material

The slightly damaged zone still contains numerous defects which favorscreep craze initiation, and it also offers the required stress level. Thiscombination results in a high probability for an early crack initiation inthat zone. The cracks then propagate into the normal material. This mecha-nism explains well the observed shorter times to failure in oxidized

pipes. In fact, MUller and Gaube report (6) that by carefully removing theoxidized layer the times to failure could be considerably improved and wereequal to those of undamaged pipes.

The above explanation correlates also very well with an observation to bediscussed later, namely that the time to fracture of LDPE pipes turns outto be shorter than the average lifetime whenever the final crack had start-ed close to the inside or outside surface of the pipe wall (see further below).

Network defectsIt has been pointed out in the introduction that solid polymers can be con-sidered as networks of chain segments interconnected by entanglements,crosslinking points and/or crystalline regions. Whereas the elastic and an-elastic properties are very often determined by the interaction of the in-dividual segments, there is no doubt that the ultimate properties depend onthe network structure (1-2).

The characteristic parameters of an amorphous network are evidently packingdensity, orientation distribution and correlation of chain segments, dis-tribution of entanglement and crosslinking points and density of chainends. These parameters fluctuate in space and time. Since each of them in-fluences the mechanicsl response of a material, there will also be a fluc-tuation of the local strength of the network. This hypothesis is generallyaccepted. Nevertheless, there is a long and winding path between this hypo-thesis, and the statement that failure of a (or a few) volume elements willgive rise to the breakdown of the whole structure.

Some of the characteristic parameters can conveniently be measured by light

scattering. Studying polycarbonate (PC) Dettenmaier and Kausch (7) foundmodest density fluctuations with a correlation length of about 90 nm whichwere ascribed in part to holes, dust particles or additives. There was noevidence for orientation correlation between chain segments even after pro-

longed annealing below T. The annealing (150 h at 130 C) did, however,

836 H. H. KAUSCH

increase the yield stress by about 30 %. With respect to further detailsconcerning methods, molecular models and relevance of conclusions, referen-ce will be made to the comprehensive review article by Wendorff (8) on thestructure of amorphous polymers.

In this section, particularly one parameter will be discussed which hasturned out to be very important and which has attracted attention almostfrom the beginning of polymer science: the role of entanglements in the

ultimate deformation of polymers. Douglas and Stoopes (9) as early as 1936,Flory (10), Haward (11), and many later researchers (12—14) have found amostly linear relationship between uniaxial macroscopic strength, b'

of a network and i/Mn:

ab a(i2 Me/Mn) (i)

with Me being the average molecular weight of a chain segment between twoadjacent entanglement points. This treatment is equivalent to consideringthe influence of molecular weight only through the 'dangling chain ends'which do not contribute to the load carrying capability of the network. Im-portant deviations from this relationship have been noted (13—14) at lowand very hiah molecular weights. Drastically simplified theoretical calcu-lations of a,,, based on Bueche's model of the ideal brittle strength of

crosslinked rubbers (12) are, however, "within the ball park' (13-14). Inthese calculations it is assumed that the ideal strength of an isotropicentanglement network is given by the force nfb it takes to break those(n) chain segments which intersect a unit area oriented perpendicularly tothe stress axis:

N 2/3 2Mab=nfb=() b(1 ---i). (2)

Whereas n and b have been quite correctly estimated, it has completelybeen neglected that the chain segments in an isotropic polymer matrix havea statistically coiled conformation. Thus, before being loaded by forceof the order of b a chain segment with end-to-end.distance <re2>"2would have to be stretched by a draw ratio

x = (<re2> / L)1/2, (3)

where L5 is the extended length of the segment. The thus defined valuesof X range from about 2 to 5 for the most common amorphous polymers (15).This means that at least locally the network must deform heavily, possiblyin an unstable manner. In order to predict reliably the ideal brittle

strength of a polymer, it is therefore necessary to know the critical con-ditions for the onset of large deformations, to identify their limits (forinstance in terms of the natural draw ratio Xat) and to ascertain therole of the entanglement network. Work to do just this has been going on inLausanne for a number of years. Some results will briefly be presented here.

Entanglements and fractureIn a recent publication (16) Kausch and Dettenmaier have pointed out thatthe role of entanglements in glassy polymers evidently must be somewhatdifferent from that in polymer melts and concentrated solutions. In parti-cular the following aspects have to be taken into consideration:

- the lifetime and strength of an entanglement at a temperature below Tgmust be considered to be very large;

- at small deformations (below craze initiation) hardly anymolecularweight effects and consequently no entanglement effects become apparent;

— at larger deformations as for instance during crazing it is the presenceof entanglements which permits the formation of stable fibrils; entan-glement coupling is also the limiting factor in the orientational exten-sion;

- the ultimate strain and energy of fracture strongly depend on the pre-sence of entanglements; it is through this dependence that part of thecalculations of the theoretical strength of an entanglement network, asmentioned before (11—14), gave not entirely unreasonable values.

Nature of defects in engineering thermoplastics 837

Starting from the idea that the gradual disentanglement of molecules con..stitutes a major mechanism in the long time and fatigue failure of polymers(2,6) Kausch and Jud in 1978 began to develop a fracture mechanics methodwith which the mechanical effects of entanglement formation could be mea-sured quantitatively. Experimental details (17—19) and results (1,17—21) ofthis method have been published elsewhere. It seems to be sufficient,therefore, to recall the principal observations.

Compact tension specimens as shown schematically in Figure 4 permit the de-termination of the fracture toughness KIQ by propagating a crack throughthe material (17—19). The formed crack will gradually heal at temperaturesabove T0 by interdiffusion of molecular coils across the interface. Thus,a certan fracture toughness Ku is reestablished in the fracture zonewhich can be determined in a second fracture experiment. Some results forPMMA 7 H are represented in Figure 5.

Fig. .4. Schematic representation of molecular coilsat fracture surfaces before crack healing; the healing

process is brought about by the reptational diffusionof molecules across the interface at T > Tg.

Fig. 5. Crack healing in Pt1MA 7 H:Fracture toughness vs

1. T = 390 K 2. T = 385 K

3.T=382K 4.T=378KThese studies and also the recent experiments of Wool and collaborators(22) have clearly shown that there is a linear proportionality between thesquare of the fracture toughness Ku of the rehealed crack zone and thesquare root of the rehealing time t. It was quite possible to explainthe observed time— and temperature-dependence of Ku by a diffusion modelof chain interpenetration (18—22). Using additional IR experiments theLausanne group also succeeded to determine the absolute value of the longi-tudinal diffusion coefficient of reptating chain molecules. For PMMA 7 H

(Mw 130 000) they obtained at 390 K a tube diffusion coefficient Dtof 32.5.10_20 m2/s. At t = 60 000 s this leads to a longitudinaldistance travelled of 200 nm which corresponds roughly to the extendedlength of an average chain molecule (253 nm).

An analysis of the microscopic fracture process at the crack tip showed(19-20) that fracture proceeded through the formation and rupture of crazes(Fig. 6). The length s of the crazed zones observed at the crack tip in-creased linearly with the chain interpenetration previously achieved during

81

puq (tLow so)'äLOMcp OCCIIL2 biesLsucigjfX iu ou MJçp wgjjsc CLS6

jOiv c,gc äkoMçp Ju bgicigjfX pejeq wgçeL4gj

çnçs qsecç M6L6 CkCf( 3LOMflJ OCCflk2 bke6LeuflgJJ7 (Eidflw6 BiG J6? MGJJ pgjeq !0U62 MJ•cp wjj cjgs puq nuqonpc6qjA couec4—M4çJJu çps 4,OUJlGk, 4,k.gcçfUG 21u4gc6 JjJ42 AL4çlou qJW4u42pG2 MJflJ JGj4uâ4u4c4jfX wo o 4LLsänjgL jocgj q42çki,pnc4ou o IJSM Guuôjew6uceflOU J2 24çJuä Cj026 ço J6 GUq 04, J6 cpi•u JJJG 2GCOUq cç COUC6LU2 J6Wfl2 pe g221u,Gq p? çpe ueMJ? ouiieq Gugudjew6uc2 o MIJJcJJ g j&de b.oboL-2622 çku24,eJ. iii çp JucGL4gcJgJ LGd4ou o 4ucowbJGçejA pjq 2bGc4wGu2fG cp4u 6Uq guq çpeu GAeucngjJ? pe LU24,GLL6 joud çJG Cpg4U' jjJfl2 çpçponäp cpiu L6bgjou gu1 11GM 6LJçUaJ6WGUç MJ.JI U6C622L.ifi 6 ouuq gç4ç 2ponjq pe uoeq çpg gccoIq4ud ço çp WOG 04, GUçUdJGWGUç 4,oLwc4ouçpe çobojoä7 uq unwpe&. o eucguäjsweuc2 Ju çps iucGucJgj djou ER.2caj2 p6q7 2eq paoe (JQ) 4MO 4,Cf2 ew ço G JwboLçguç MJçp LGägq ço

wojeciije2 cu 4,OLW g copeksuçj? euçudjeq bX24cj UGçMOL<çpgu çpgç o çpe c.çJcj wojGcnjgL. MeJäpç WC. OujX gç 211CJ g jGUäçp flJ6

çpgç eI° LJ262 uoçpj MJGU çpe cpgiu2 GgCp g jGuaçp MIJ4CP 42 JäGLM6L.6 E(J) 42 Aofluä,2 woqnjri uq / boi.22ou12 I 42 MGJJ Op2Ga\6q

010 = E(ic)(r-) ()

JG A4a4u wgçsLJJ p7:] JjJG 4,LCflLG GueLdA 010 ejçeq ço çpe LgCçnG çondpue22 K'0 otaju 2beciweu2 i2 gjo LGAGgJGq pX PG exbeuweucgj qgçg bjoççeq 4u E4n6jpe 2çLouä iu4,jneuce o, eucuajeweuc2 ou çpe gccne 6u6LäX 010 04, A4L-

çjgjj? pejeq bNNV H (ow soYI1U46L6UC6 wiCL0äLbp 04, g ouç Ju bg,-

LaC LOU !U baL4!aff1 peaq bI1w1v

W112 JgAG 4uAojAGq JG 4,0Lwç40u 04, eucgudjeweuc2 (jo)ous cu 2gAefX coucjnqe çpgç çpe iuceLq44,4,n24ou 04, c}Jiu2 qclL4uä .epegj4uacguajeq Cpiu2 4,LOW g wgjx LGdfliLG2 4,LgCçIILG eueLd4e2 04, 0,1 0 1 ') W_5iuä g2eq ou JG exbeLieuce () çpgç çpe 24wbJe brijj 0ii 04, uou—eu—

Mp4cp 4u bNWV wortuçeq ço 13 ço 380 () w_ qebeuq4ud ou çpe qGdLG6 04, pegj-LG2bOuqGq AGJ.? M6jJ 0 JG 4,LCflLG GU6LA I4 04, çpe ieiejeq 2Wbj62pejiud' jJG GuGLaX 220C4çGq M4çp çps 4,OLWgç4Ou 04, J6 cgeq 0U62 COL-

838 H' H' RVH2CH

E8 aiIqorrnrrJi g ssrdgris n 2is1b to cuisl4

o m2rnscb9m 91U 29VLI6n6 9fl0 n9IiW bnm n eniod ed bluoda 2Im owi eaeriT

-'l9iIlr 91t1 nI .2n9mb9q2 beI6eIi VjeierqmoDni n noLi6m1o [hd bn6 951D -ears n5IIJ 2in9m9Fpn5n9 19W9 916 919fi 2fl91fl 9q2 IiDLJ2 O nO19l FD bemuaa6 9d nsi1t taum I .2bn9 niz siJi oi seofD bnuo sis er1i bns ,91911W

211151i3 berpnsine rwen erli 'f1siceqee noiisffi'TdH o aae3olq etii n1 it 2In9m91pntn9 seoth pnieool yjFbubblp 916 y9di isdi bns bseaeiia '1iiii 9,5

to noiiu[oa2b eislqmoD erii 'jinebvJ .ebn9 fltblb 9th oJ 920F3 915 rbrrlw bse[ 1Fw [hdit 5 0 n01i392-22013 lsIuz)ri'lsq 5 111 2in9m91n6in9 W911 911i

-X9 'fifl9fn9Vfl03 bluow noriqnlu22s evods eriT .1iidi isrii o 9lIJiquI erti oi bns norismioeb 1hd o ineixe erli n9ewied noLIsrelloD bevisedo erli nsFq

.(OS@I) 11i1591191

-911 915 291U39101T1 fl99Wi9d einemerrisine isrii bse ed vsm i no2u13noo nI 11911i i119V9'lq bns 2i1191flp92 1199Wi9d 19t2n51i 2291i2 911Z1 9xrUdsie ozl '152293

anoi-isoirqqs nrised bsol 1o b92u 23ii2sIq 92011i nI .nOi515q92 bqsi -sic oi bebeen nSIIi 91u39101n ieq ain9ru9IgnSin9 910111 9d (,1Ts1ene ffw 91911i

—n9b ineme1risine O 2nozt6uiDurt rSDOr rrsm .192n51i 2291i2 bqsi ssird to iedmun 9tii eI9V9WOII I .93119up92n03 011 O 9d Oi 111992 91O919rti i2

smurov ed rrrw 9,91ii nsrii bease'ioeb 'Fdsion ar eruDerom ieq ainemetpnsine eesioeb A .aeiia i3eeb eiuiiiano3 iizthiw aeru3erom berquoonu riiw cinemsre 10 wor e110f22f32 1115113 tipuo'iiIi 1330 1163 nOiiSliIl9DnOD in9ffl91n5in9 to

.(aS) @nftsof euiis ruoi11i '10 (a1aS) pnftsof erni—nof

noii,iihil ss'i3 'd bev'ieado rine3e1 25 (II 2ex513) 29Th13 3I2flhiflf O noisiiiiot eriT

11z1rlw ifdsian 115 ceiuiiicno3 \r,ser3 (OC-S) rl3aus)I bns esmneiieQ erqmsa erli oi beio'r"1i291 ion a bns aeiiieqoiq howien erli no ebneq9b

beisin 91S 29X513 3renrlixs 10 Fsnorinsvno3 erli oi beeoqqo cA .eDshue nks13 to 2e351i 'd io\bns 2er3i1sq ieub 10 ewsr to boor1iodripen erii n

.(8 .pH) 2e35hu2 nembeqa elli is vjrsrineiseiq 2ine5

-ine'iosq n (I aexsio) 29X513 1snoinsvno3 .8 .pH 30 II is berklielie '1ineupeedue eisnodls3'(,foq be

. S = 3 oi

abser II 2esi o (01 bns .apr9) iedrnun e,s1 'iev edi Po noisin'r 91ff bns 9V'13 nsia-aeeiia nr'ieenrne erli o qo'ib s 10 norx9flnr 116 oi —nomeb need asd ii .bertUneb ed vLI2se vLsm noisiin'r 9X513 e1oe1erti

—xe ns is 31 beinei'ionu 1rsniho n beisiitnr is II aexs,o isrli beisiia 11iw rrew ieriis, aeeis eursv airiT .('cS) 1.S =

''A iuods o oiisi IlOi2nei .iS11A ainioq inernerpnsine neewied anisrb o 'iiridianeixe mumixsm erii

'd ietliie ans11o o inemefpnsineaib isili bernuces ed ieum ii eioeie11T II 9513 10I eroi insi'ioqmi 116 a'srq eiuiqui ni5113 'd ,o esqqra ni5113

.i1iwoi bns noiismiot

HDJA){ .11 .H

n (II 29613) 29S5'lD Di2nhin bn6 I 29S513 .2 .II 30 211 i6 b9iii9'1I2 'jin9up92du2 9i611od163\(1oq b9in9hoelq

= 3 Oi

'm11kd eth o ciq61o13m n01i3919 2nnn2 .01 .2H .9t6n0d163y10q betneHo ni II 296lD O 91Ut3U1i2

2)11OWt9n be)1nU22ol3 111 91Ui3U1J2O131M Po 29Zti2n9b )1nrreeoI3 9V611 2flf291 2nrfl92Onr1911 be)112aol3 rduoior1T

oi 1 O 2)1111122013 n99Wi9d 2e3n6f2b 92619V6 .e.r ThiD O1•S oi 1.0 iuodb —viiq eitt ri 2noii6ui3u1 i61it 910919r1i i39qX9 oi 2611 9110 io'iq A .mn S

6 0 2n9ffl919 ernutov 9VIOVnr FIrw )llOWi9n )1n1122o13 911i to 91UJ3U112 1632 -3u1t2 vjevti6IiIn6up 9nrm1919b oi i[u31 ft i 'FIn9bv3 161m2 .hb 2hii 2t39191 911Ji619fl1 th1911U3 9111 .921161—11111 911i (11 2nOi6r16V 9ltJi

01 eVLfl2n9b )1n122o13 ti11 o 2flO91 l6Iubon 110 2i10q91 916 9191ff .ifu3 2 oi 8 2913ri16q 1161112 no 19i9m6rb n mn 000 01 oi 000 8 o216 iud Inn oa oi

1163iiqo o bn 12nn6'12 pni'iub ieilio 11369 i25q wofl 113111W 9i2 111 1011

b91163 02 911:1 O 29t2 9111 .(CC-IC) 19i9fl16b n mm 1 291ub0n 91d6n1932tb bn6 B 2ulubom e'2nuoY 1t 11i2n91i2 mo'il b9161u3153 916 3 2w61 32ni1if1

:n0i6up9 I1iHIO 9111 oi 2nb1o336 'y '19119 936hU2

()

.m 0 iuod6 9d oi

163i3eriioq'ri 10 16Ui36 92911:1 0 93n9u1n eili nO W91V benu 011 2 919111

.16 is 3V0M .2ni291 b91u3 0 lOiV6fl9d 91Ui361 911i 110 291it161U2911t

-91 VL219fl9 1631:1113 erii bn6 noLi6m1o9b 3U261q 16301 erli :1611:1 9i6i2 (IC) b9i61911o3 916 bn6 W01t l6IubOnl9in1 d ben1mieieb 916 9i61 92691

.9XF2 9lLJbOn itl1w

i9 n621oM .9263 16139q2 6 ed 11 ew ym i6clw o eb6m 9d 0216 b1uo112 nolineM 191ii61 0 mio 9Iii n1 2111291 'x0q9 b91u3-enm6 '1ii11f bs1buie (SC) .1 6

2n1bn6d 169112 e2nis613 wo1 2922612 92911i n1 b9vl92do e111 .an9mi39q2 nhii

jocgjj? npecçeq ço js qsouuçouLG uoç qsçcçpje 411 2pOiç çJ•W6 jogquä G22 (11JJ622 4J6 646CflA6 O1J6 2ugw6jX çpgç CIseb ciss IJI1CJ6J Wfl2 pG COU46L6 g2 ueçMo qsscç M4Cppgo couijjw çpe ççsweuç wgqs peo () uq 4U çpe bLen4or1 2GcflolJa2gwbJ62 äGU61JjX pio<e MJ6UGA6L çp U6C( pgq 16Cp6q çpe q6ecc pe—

ejouägc4ou o4, Jjjw coucgjuiuä q646Cç M2 wncp wgjje. 241JC6 çpe uecquâcouc6uçL.çJou O1 4JM2 uq JUCJI12JOU2 011 çpe OflJG HOMGA6 1J6 66CflA6floU 6M66U 7uejq 2L62 guq çwe ço 4g1JnL6 ou çpe ous puq uq quq uqiow çpe beeJ4ud2 ju çpee exbeu.weuc2a jpee M2 bLgcçJcgjj? uo cOLL6jg—beoLw6q 4E,u24J6 qLM4uâ uq çç4c jogq4uä exbeLJweuç2 Ou 2wbj62 cnç10 çeç JG 4UtJII6UC6 04 çpe 0p261AGq 4LL6änJgLç462 Jj11flJ6L fiG gncpo

3sf)

MpJCp e2Cbeq wixiud Mfl cgkpou pjc (4j.owEJä i1 OUG oi,, gporiç 3 ww jeuäçp ju U bibe

Lbiq tgijflJ,62 ()fJJ2 LGäJOU çpgç guq M4uceLäeL2c oniq fiG olia4u2 04, fiG W024,iI2ç çoGxbG!JGuce gu eieuçnj ueçMolJ< qeäLqc4ou ic M2 bLGCJ2GfX ju

2GCçJOU çpe JUUGL. MJJ2 o exçnqeq bJbG2 qoMu ço g qebcp o O WW gIG flJ6ge wo2cj7 ie2bou24pje 4,oL op2GLAç4ou' V2 q42Cfl226q u çpe bLeceeq4uatJflGUCG CLgC iuJJgçiou 4ç 266W2 ço pe woe bLoppJG çpgç U6M0L$ G4,6C2MJçpJU fiG bibe Mgjj qne ço 4,LOGu—4u cGu24oue (266 uexç eeCnou) w7 4u—çpe CGUçGL (P = oo p) vjcponäp fiG q44,4,sLeuCe2 411 2L622 q42çL4pnç4ouu4cnqe wjjeJ (çP = stü p) çpu çpoe MGLG çpe CLgC pgq ok4ä4uçGq 411CJ026 ço fiG 4111161 01 0(1GL 2nu,,gCG 04, fIG bibs gjj M6IG gu IGk 04, wgä-çpg fiG J44,GçJWG2 çP 01,, fiO2G b4be2 MGLG fiG CIGGb CIgG2 pgq 2çgIçeqfiG iUeanjgIiçie2 MGL6 woke 4,kedneuf ic M2 bLç4CnJLjA uoeq pA çpewCS çps i,,iugj Cleeb CL!G2 ok.iäiucGq 411 fJO2G LG2 04, fiG b4bG Mjj M6LG(Hä ii) uq bLJj6j 2C1igçJou2 jpeA Cjgiw çpç Miçp g CGL4giu bLG4,eLGu—uoç wixeq M4çp C1pou pjc< ju çpe 4,0UJJ 04, boiuç2 JiuGgI. WgJ$2 00(24,gijnl.G çP o çpe bibG2 jpG JuGdnJgkicie2 çpeA oriuq MGLG CJGJ S011G2gpje ço cou.ejçsq gffl) 4,IGdrlsuCA uq bo24ç4ou 01,, G4,GC2 M4çp cwe çobesj4uä 04, ço WW fi4C(u622 B' 4u2bGCf4uä çp42 4,4JW fleA M61G

Mp4Cp bGIwiçceq ço çku24,oLw fiG euçi&e bibe Mgjj iuço g COuç4uOfl2 fi4u(rDbE) b4be2 çocwgAsL uq MJ.uG1äGL2c (j) qeAGJobeq g CIlçfluâ çGCpuJdflG111 oLqGL. ço qGuc44,A 2A2cewgc4CgJJA cue ugçru.Guq bo24c4ou 04, 64,GC2 411

4,GM IGWL(2 CO11CGL.u4uä fiG joCi0u 04,, 4,JM2 uq 4uCJfl240u2 M4JJ pG wqG'wiCko2CobiCgj iuAe2çJäçiou2 g2 ço fIG 0LJälu 01,, boj'XwGL 4,4JflLG2 g

COuCGuf.giou G4,,4,GCç2 04, CLgC(2 01.. bgLç4cJG2 uq 011 fiG UIIWGI0112 GJGCçI.OUj42 gLçiCjG Cgu uoç ççswbç ço LGAJGM çpe gpnuqguc JiçGkgçnkG 011 2I622

42 jG22 JGgL) bLçiCnJijA II,, fIG 4,JM2 01.. iuCjfl2iOu2 o1iaJugjJA gIG

4,JI2ç boiuç (01.iäiu) Cu ä6uGLJjA P6 qGcGkw4usq fluWpJä0fl2jA fiG 2GC0Uuq kGqIICGq fiG 2çLGUäfJ 01.. flW6 ço 4,4J11LG 04, fiG 2bGCJWGu GLG2 fIG24011 M4C gncowciCgJJA MJJJ pG pjweq o1.. pgAI•uä 21G fiG i.uj CLgCVJW02 uA L4jG 4,LgCçnLG 2flL4,CG Cougiu2 çLC62 04, g 4,JM 01. gu 4uCJfl-

EYM2 VWD IVCrfl2IOLi2

iuä (33)4,gAoL.6q 4upowodGuGon2 qeowgou 411 fIG 4,01W 04, CLgiud uq\oi. 2p6g1. pguq-M0L<2 4,A01.Gq pOWOäGUGOI12 bj2ç4C qG4,0JWç4Ou MGLG2 jG22 bGL.4,GCç UGçM0J(2uq C1..gCiuâ IPGA COuCjiiqGq çpgç bGL4GCcJA powoäGueon2 CL.o22J4u(sq uGç—

4f1L6 O G6CC2 nj GUTUG6LTU CJJGWOBJ21TG2

J0 fiG cp41J gxi2'JjJn2 g âJ.OMJUä ck.gc( bi,e euçijfX 2Gbkçe2 cpi,ugc 41.1 çp. qJ1.ecc4ou u1bç1nG occn ju gxigj q,Gcç4ou bGLbGuq4cnjgLfX(bG1pb2 p7 g gco 0i To)' vjcpondiiçpe jLdGeç 2L622 cowbouGu 2çflJecç4ou' u çp qpecc4ou i,icçn 2cLGuäcp pg cou4q6LpJ? 4ucLGg2Gqqsowq ouG cou242ç2 o WOJGCflJG2 p4dpfX osusq u c cnweisuj q-HDbE b4bs GbiG26uçGq pi EiänLG j äi,ie gu GXC6JJ6U exgwbje' jjis peAJJX2gwbjG2 2OM 2JäU2 o guJ2afLob1 (piLG4,Liud6uc6 exçeuç 04, 2pLJU(dG)' jpGCflj62 uq o guk oeueq boJ?wJG& wçLJx' gu? 4ulecc1ou wojqeq o exculqGq2fflCflLG gui2oçLobX ou oi çp cpgLgcj6Li2çic bLobeLfle2 04, Cpiu WOje-OLGucç4ou q42cL4pIlçl,ou2

GUç4kGJ? uGäjJdJpJe'sxçsujjA bbJGq AJflG2 04, O (3 bg fiG 4,LOGU-JU ç6U2iOU2 JG U040flç24q6 uq g GU24JG cowbouGuç 04, +1 gç ç 4u24q6' owbsq M4fJ çpEOL g HDbE bibG pG pgq qGçsuni,uGq (cowbLG22nG) 2JG22G2 04, -3 gç çM MjJ cpJCuG22M L4fl22GJ unWpGL.' boJ,220u2 çi,oE(c) Aofluä,2 woqnjri g çiwe g4,G. coojiuö qoMuX C064,4,4C4GIJç 04, çpGi.wj GxbglJ2iou2 q42çucG 4,LOW 4u24q6 MgJJ

MJçp:

=1

(WC) () - () eIilfl 1 2 5

JJJJuJ2 (3p) q6kJAGq fiG GUJJj C4LCnUJ4,GJG1JflJ 2ç1,G22G2 ag(2) g2:cowb1.,G22JAG gç çpe 0flç24qG wjj guq cGu2ijG gç çp Mgjj 4u24qG'GxbGLiwGuç (32)' iu bJbG2 çpGA äGuGLgJJA 2OM g bgLgpojic q42f4pr1ç4ouuq gpojnc& AJfl6 0l 211Cr) cGu24ou2 cu pG qG4GLWJuGq 4,1.0W g j'AG1.. LGW0AJuGOI12JA uq G4JG1. puq 01. C0uçl,u2 4,k0Gu-4u çGu24ou2' JJJG q42ç&4pflflouçGwbGLgçclLG (f oi i") ço g J0MGJ GWbGLgçnIG 1C qG4,0J,1J12 uou-pOWOäG-yuA Ai2coGJ2çic poqA coojeq 4,LOW onc24qG 4,L0W 0AG 4ç2 20j4q44,4cC40uuçG,uJ cGu24ou2

I = 80 C = 0'2 P)

Ha' 5Q oncc1jG 4,4JIG ot HDbE bJbG (aA = bg

Ju fiO2G C2G2 42 iuicigcGq pA UGM0L G4,GC2 OL 4,jgM2'(pL4ccJG 4,Lcçn1G) oL JouaGl. çJ,WG2 (cLGGb cg4uä)) GC112G fIGoçq pgj fiG gjjiuä 2Gc1JOu wg'X G bGL4,GcjA 2g4,G gç pJapGk. 2çkG22G2äGuGLgJJA <G2 bJcG gç çpG 2WJJG2 cJ022-2Gcc40u (E4d' is)' ic poiijq GqG4,GCç' j112 fiG qnccJJ6 J,JflLG 04, nuigxigjjA 01.. pixijjA joqq 2WbjG2y 2JUIbJG ALigçJou Ju 2gwbjG cLo22—2Gcc4ou cu çnu onç ço G g qGC424AG8G0WGLA

OIHEE� b0DflCII0W EV1ED DEEECI2

cowbLGpGu24AGjA nuq. Iob4c J/ 04, çp42 WGGç4uä'guq 112G 04, bojAwGL cowbo2içG2' 1p42 bLopjGw pg2 cPGLG4,OLGa PGGU qjç M4fJpJGuqJ,ud cowbouGuç2 o1.. .LJjJGL2 42 04, A4J 4wboLcgucG ço fIG 4,pL4Cç40uEn4qGuc fIG dnG2çiou ou çpG iu4,jnGucG 04, 4ucj11240u2 gqqeq 4u fiG 4,OLW 04,

H' H' KVTI2CH

E8 a3baBIqomscfi rsriiris 111 ts1sb lo 91L11SV1

9r111'l9bnu fno b6'lD 9rlt pnpbhd 1hdit 911i) mr1i p11b169'id o bien .(in9m9zIie 2riZt

een1 bfsW tnfihoqmr i2orfl 9rti 29IuZIif2nOD b'mwioi1i'1i2 bn6 bq61 91qrfl2 nibTeW 1611ñ ouT .etnemele noitu1ianoD Dri26Iqorrrlerii ni1dme226 io eupnrizet

'irerli 'd 161'ieism 1116fl1 91i mo'u '1Th913 29V1921fl91i Ii2hJ1Ti21b y,U2onl eblew -n9 bns) ioiir1 lsmieuii (IJ: bns CI .epH 992 19d1n53 e2bS9d 1ieequ) iirio

—9neq'iein '161u3910m bum 11Oit5ifl9i'1O O 9t5t2 e(29fl5f13 ls3o1ouq1om nhie ifl9ffl9p11iqmr erli 'd b9:fselo 291111 blew 9201ii 915 2uoivdO 22ef 119fl0 .noUs'ii .bseti 11OI2LJIiX9 115 111 19b1q2 eci bnuo'is 'io blorn s nr 29225111 nwo1I o

-919 Dr2Bd 2hlT .15ii119229 21 noiis'1i9n9q'19n 'lSlUDeloIIl boor nis 9'19tIT

eno1i5m1ot Jn9m9Ipnsin9 ripuoirii rIineiia o qu-blhid 911i riib1ew 0 ill9fll -X9 911102 inoq iri iA .1101i392 19r1m9 115 n 16i9b ri be22u32b ne9d sri

.bebbs 9d 1lw 1in9'1i2 (en1) blew rio 21u291 lsinerniieq

eqq ee'irU oJ beblew-flud iiio-T 9fl9I'q0'Tq'l0 .CI .prR -iifl9bi owi ru bobbin fl0ii39jj1 25W inrot.—T 9ILi 2flOii392

.beblew '1i119up92du2 29V1511 163

-r39q2 evods O 1911103 bnsu1-i1o1 1o qu 92013 .E .prR niiiw belluDDo 91u15t (3° d 15d ') Zt29i 19it5 (19ffl

eili nr '1d22oq 1e2i iiioL-T 9th Io 91151q bI9w erli

.1911103 V19V

219fl'1O3 25 113U2 291ui59 1s3qi thiw bbori s b9ve3no3 25th (SC) biowsi3 21163 eth isthw iuo behm3 25th bus .oie 2b19w wol eflOiD92 fl 29115th3

9th ed oi iuo berriui 291111 blew orU 11 thiw 2i29i 2111 nI .'enbloni—obueeq" .ioeteb eldsion i2orn

844 H. H. KAUSCH

Moslé et al. (37) molded PC plates with two central holes. By simply chang-ing the stretching direction of these plates in subsequent tensile teststhey could either load their specimens parallel (a) or perpendicular (b) tothe weld lines. Their principal observation was that the yield stresseswere more or less identical within a range of molding temperatures v3ry5ngbetween 300 and 340 C. The post-yield strains, however, were drasticallysmaller in case b. The rapid disintegration of the weld line material dur-ing plastic deformation is also underlined by the lower rupture forces (a/b= 2 200/1 100). Once again it has been demonstratedthat strain hardening,a vital property for engineering thermoplastics, can only occur if the mo-lecules are reliably interconnected so that they do not disentangle at theonset of plastic deformation. .

Acknowledgement - The author would like to thank Drs. W. MUllerand E. Gaube, Frankfurt/M.-HOchst, for having provided some ofthe test specimens and for valuable discussions.

REFERENCES

1. H.H. Kausch, IUPAC MACROMOLECULES, H. Benoit & P. Rempp Eds., PergamonPress, Oxford—New York, 211 (1982).

2. H.H. Kausch, Polymer Fracture, Springer, Berlin-Heidelberg-New York(1978).

3. B. Rânby, J.F. Rabek, Photodegradation, Photo-oxidation andPhotostabilization of Polymers, Wiley, London-New York (1975).

4. U.W. Gedde, On the necking and fracture behaviour of polyethylene, PhDThesis, Royal Institute Of Technology, Stockholm (1980).

5. U.W. Gedde, B. Terselius, H.-F. Jansson, Polymer Testing 2, 85 (1981)6. W. MUller, E. Gaube, Kunststoffe 72, 297 (1982).

—7. M. Dettenmaier, H.H. Kausch, Coll. & Polym. Sci. 259, 209 (1981).8. J.H. Wendorff, Polymer 23, 543 (1982).

—9. S.D. Douglas, W.N. Stoop,s, md. Eng. Chem. 28, 1152 (1936).10. P.J. Flory, J. Am. Chem. Soc. 67, 2048 (1945r11. R.N. Haward, The Strength oTPTstics and Glass, Cleaver—Hume Press,

London, Chap. II (1949).12. F. Bueche, Rubber Chem. Tech. 32, 1269 (1959)13. B.H. Bersted, J. Appl. Polym. SEi. 24, 37 (1979).14. D.T. Turner, Polymer 23, 626 (1982).15. E.J. Kramer, Lecture series on Polymer Micromechanics, Swiss Federal

Institute of Technology, Department of Materials, Lausanne (1982).16. H.H. Kausch, M. Dettenmaier, Coll. & Polym Sci. 260, 120 (1982).17. K. Jud, H.H. Kausch, Polymer Bulletin 1, 69Tfr97J.18. K. Jud, H.H. Kausch, J.G. Williams, J. Mater. Sci. 16, 204 (1981).19. K. Jud, PhD Thesis 413, Swiss Federal Institute öf Technology, Depart-

ment of Materials, Lausanne (1981).20. L. K5ncz5l, W. Dbil, H.H. Kausch, K. Jud, Kunststoffe 72, 46 (1982).21. T.Q. Nguyen, H.H. Kausch, K. Jud, M. Dettenmaier, PolyiiIr 23, 1305

(1982).—

22. R. Wool, J. Polym. Sci., Polym. Letters Ed. 20, 7 (1982).23. H.H. Kausch, K. Jud, 5th mt. Conf. on Deformation, Yield and Fracture

of Polymers, Cambridge, 29.3.-1.4.1982, p. 9.1.Plastics & Rubber Processing & Appl. 2, 265 (1982).

24. E.J. Kramer, J. Mater. Sci. 14, 1381 T1978).

25. R.N. Haward, H.E. Daniels, L.R.G., Treloar, J. Polym. Sci., Polym.Phys. Ed. 16, 1169 (1978).

26. P.E. Bretz7R.W. Hertzberg, J.A. Manson, J. Appl. Polym. Sci. 27, 1707(1982).

—27. M. Dettenmaier, H.H. Kausch, Polymer 21, 1232 (1980).28. H.H. Kausch, M. Dettenmaier, Polymer BUlletin 3, 565 (1980).29. M. Dettenmaier, H.H. Kausch, Polymer Bulletin , 571 (1980).30. M. Dettenmaier, H.H. Kausch, Coll. & Polym. ScT. 259, 937 (1981).31. J. Mijovic, J.A. Kovtsky, Polymer 20, 1095 (1975)32. R.J. Morgan, in: Developments in RTnforced Plastics—i, Chapter 7, G.

Pritchard Ed., Applied Science Publishers, London (1980).33. R.J. Morgan, E.T. Mones, W.J. Steele, Polymer 23, 295 (1982).34. P. Stockmayer, S. Wintergerst, 3R internationaE2O, 274 (1981).35. J.G. Williams, Plastics & Rubber Processing & ApT. 1, 369 (1982).36. R.J. Crawford, 5th Int. Conf. on Deformation, Yield md Fracture of

Polymers, Cambridge, 29.3.-i.4.1982, p. 38.1.37. H.G. Moslé, R.M. Criens, Kunststoffe 72, 222 (1982).